class: center, middle, inverse, title-slide # Modern brains and bones: ## Genomic analysis of derived <em>Homo sapiens</em> traits ### Alejandro G. Muñoz Andirkó ### Cognitive Science and Language PhD. program ### 26-07-2021 --- class: rutgers, inverse, center, middle # Chapter 1 ### Human evolution in light of paleogenomics --- # The paradox of *sapiens* uniqueness .pull-left[ ### More and more **similarities** between Neanderthals and us - Several admixture events with both Neanderthals and Denisovans - Similar levels of cultural complexity: - Lithics - Pigment use - Cave art - Body ornamentation ] .pull-right[  .center[(Nubian Levallois technology associated with Neanderthals, Blinkhorn et al. 2021)] ] ??? As a push against Great Leap Forward --- #The paradox of *sapiens* uniqueness .pull-left[  (Credit: Phillip Gunz, CC BY-NC-ND 4.0) ] .pull-right[ ### ... Yet there are important **differences** - Anatomical differences in endocasts: diverging parietal, cerebellar and (possibly) subcortical morphology - The existence of deserts of introgression that resisted admixture - Regions under positive selection] --- class: center # How do we reconcile these two views? ### Integrating paleogenomic data with large-scale -omics databases that take advantage of **current variation** in modern populations -- ### Using clinical data from **pathologies** known to be caused by **targeted, restricted** genetic disruption, and affecting phenotypes that have changed over human evolution -- ### ... all while taking into account that this data can't be really understood without a **temporal dimension** of variation ??? Each one of these approaches corresponds to a chapter --- class: rutgers, inverse, center, middle # Chapter 2 ### Effects of high-frequency *Homo sapiens*-specific variants in brain gene expression --- # Questions - Often, too much emphasis is put on **missense** variants, but is there a role for **regulation** in human evolution, as predicted? - Can we infer **which tissues and genomic regions** are particularly affected by *Homo sapiens*-specific variation in gene expression? -- # How can we adress these questions? - GTEx single-tissue expression quantitative trait loci (**eQTLs**) across **15 brain structures**. - Filtered through **Kuhlwilm & Boeckx (2019)**, an exhaustive catalog of variants differentiating modern humans from Neanderthals and Denisovans. - Regions identified as under **positive selection** in two independent studies. --- # Results summary ## Permutation tests High-frequency eQTLs are **overrepresented** in genomic regions under positive selection.  --- # Results summary ## Tissue-specificity - In terms of eQTL variance across tissues, the **pituitary** and **cerebellum** accumulate more eQTLs than expected by chance (against control sets). - Controls: non-derived & non high-frequency eQTLs. ## Others - No directional skewness. - Overrepresentation of specific functional categories in high frequency derived eQTLs relative to control: **NMD**, **5’-UTR**, **non coding transcript** variants. --- class: rutgers, inverse, center, middle # Chapter 3 ### BAZ1B, Williams-Beuren syndrome and the evolution of the human face --- # Williams-Beuren syndrome - **Neurocristopathy** caused by the deletion of an specific region in chromosome 7. - Typically results in a characteristic **craniofacial morphology** reminiscent of *Homo sapiens*-derived traits: eg., retraction of the lower and mid face. - **Sociocognitive changes** as well: exaggerated social tendencies, increased trust and friendliness, reduced reactive aggression, relatively intact language. -- - All of this has led to parallels between Williams-Beuren and the "domestication syndrome", and a potential way to test the **"self-domestication" hypothesis** [1]. - Along with **Giuseppe Testa's lab**, we set to provide the first experimental validation of the hypothesis through the effects of **BAZ1B**. .footnote[ [1]: Though my current stance on the term and the hypothesis overall has changed, as stated in the introduction.] --- # Cohort and experimental design  --- # BAZ1B-bound genes  --- class: center # Summary of results ### The most stringent category of genes following BAZ1B dosage is tied to an **excess of regulatory changes** in *sapiens*. ### Neanderthal/Denisovan enrichment in **inversely correlated** gene lists. ### *Homo sapiens*-specific enrichment in regulatory variants **directly correlated** with BAZ1B. --- class: inverse, center, middle # Chapter 4 ### Temporal mapping of *Homo sapiens* variants --- class: center # Questions ### Most human evolution studies disregard the temporal dimension of their results. ### Simulation-based age estimates are usually dependent on unkown demographic parameters. ### Can we derive a temporal dimension of variants in an evolutionary scale? --- # Albers and McVean (2020) - Non-parametric method - theoretically agnostic to demography. - 45 million age variants estimated. - Hidden Markov model + pairwise coalescent-based TMRCA estimation, using recombination, mutation and joint-model clocks. # Downstream analysis - **ExPecto**: deep learning, variant-based *in silico* predictor of expression variation [1]. .footnote[ [1]: Through a convolutional neural network train on "cell-type-specific model for 2,002 genome-wide histone marks, TF-binding and chromatin accessibility profiles" ] - **Time-sensitive GO** enrichment analysis. - **Case example**: *BAZ1B* --- # Summary of results ## Including control  --- class: center # Summary of results ## Expecto: We found genes related to cerebellar Purkinje neurons (0-60k), glutamate system (300-500k), dopamine system (500-800k) ##GO enrichment: Bone categories found only in 0-300k, consistent with our vision of the archaeological record. ## *BAZ1B*: At the same time, older variation in variants potentially affecting the regulatory network of *BAZ1B* than expected initially. --- # Conclusion ### Each of these three studies **complement each other** in perspective, and ask similar questions at heart - that is, trying to determine what really differentiates *Homo sapiens* from Neanderthals and Denisovans. -- ### While the answers in this thesis are partial, they aim at all points to derive information from open data while **doing justice to the increasing complexity of our species' history**. -- ### An **interdisciplinary view** of human evolution is key to derive conclusions from all dimensions of the paleogenetic data. --- class: center, middle # Thank you! --- class: inverse, center, middle, remark # Extra slides --- # Q&A: The "domestication syndrome" .pull-left[ ## Promises - Comparative genomics across different species models - A localized, testable hypothesis, centered on neural crest cells - An expansion of the term of "domestication" as a biological process, rather than a cultural practice ] -- .pull-right[ ## and practical problems - Not consistent across species (except for reduced reactive aggression) - Unknown relationship with other components of domestication (eg., glutamatergic system) - Pleiotropy and the overall general role of neural crest cells in early development ] --- # Q&A: Divisions in temporal mapping project